Melatonin effects on luteinizing hormone in postmenopausal women: a pilot clinical trial NCT00288262
© Kripke et al; licensee BioMed Central Ltd. 2006
Received: 07 February 2006
Accepted: 16 May 2006
Published: 16 May 2006
In many mammals, the duration of the nocturnal melatonin elevation regulates seasonal changes in reproductive hormones such as luteinizing hormone (LH). Melatonin's effects on human reproductive endocrinology are uncertain. It is thought that the same hypothalamic pulse generator may both trigger the pulsatile release of GnRH and LH and also cause hot flashes. Thus, if melatonin suppressed this pulse generator in postmenopausal women, it might moderate hot flashes. This clinical trial tested the hypothesis that melatonin could suppress LH and relieve hot flashes.
Twenty postmenopausal women troubled by hot flashes underwent one week of baseline observation followed by 4 weeks of a randomized controlled trial of melatonin or matched placebo. The three randomized treatments were melatonin 0.5 mg 2.5–3 hours before bedtime, melatonin 0.5 mg upon morning awakening, or placebo capsules. Twelve of the women were admitted to the GCRC at baseline and at the end of randomized treatment for 24-hour sampling of blood for LH. Morning urine samples were collected twice weekly to measure LH excretion. Subjective responses measured throughout baseline and treatment included sleep and hot flash logs, the CESD and QIDS depression self-ratings, and the SAFTEE physical symptom inventory.
Urinary LH tended to increase from baseline to the end of treatment. Contrasts among the 3 randomized groups were statistically marginal, but there was relative suppression combining the groups given melatonin as contrasted to the placebo group (p < 0.01 one-tailed, Mann-Whitney U = 14.) Similar but not significant results were seen in blood LH. There were no significant contrasts among groups in hot flashes, sleep, depression, or side-effect measures and no significant adverse effects of any sort.
The data are consistent with the hypothesis that melatonin suppresses LH in postmenopausal women. An effect related to the duration of nocturnal melatonin elevation is suggested. Effects of melatonin on reproductive endocrinology should be studied further in younger women and in men. Larger studies of melatonin effects on postmenopausal symptoms would be worthwhile.
In vertebrates, melatonin serves as a hormonal signal of night, and consequently, as a signal of the night's duration through the changing seasons. In a number of mammalian species, the duration of the nocturnal melatonin elevation regulates seasonal changes in reproductive hormones such as luteinizing hormone (LH), follicle-stimulating hormone (FSH), and prolactin [1–4].
Whether melatonin regulates reproductive endocrinology in humans has been uncertain. On the one hand, melatonin was tested with some success as a contraceptive . In a controlled trial, 3 mg. melatonin at bedtime evidently suppressed LH in younger women as compared to placebo, but not among women of menopausal age . Several conditions of inadequate gonadal function or amenorrhea may be associated with enhanced endogenous melatonin production, but not all reports have been consistent [1, 7–11]. A prolactin-stimulating effect was noted in women given melatonin 2 mg at 16:00 and 20:00, but there was no significant effect on LH or TSH . Melatonin decreased LH in men in one study,  and melatonin impaired sperm production in two young men .
Wehr and colleagues subjected experimental subjects to 8 and 14 hours of complete darkness, without attempting to regulate light levels when subjects were out of bed. Dramatic effects of long nights on melatonin, cortisol, and prolactin were demonstrated . The duration of melatonin secretion was prolonged . The duration of the nocturnal prolactin elevation increased, but it was evident from the mean plots that the peak values were higher during short nights, and therefore, it is unclear what the effect of long nights were on total daily prolactin secretion . There were no significant changes in LH from baseline constant routine measurement to measurement after 4 weeks of 14-hour nights; however, since only 3 blood samples were obtained on each occasion, and LH is both highly pulsatile and rather circadian, the sampling did not conclusively address the possibility of effects on LH . Moreover, a man and a woman subjected to 14-hour nights for 15 weeks experienced a disturbing loss of pituitary volume. In that one man, LH and FSH increased during a period of normal nights that followed 15 weeks of long nights, suggesting that secretion of these hormones had been suppressed during the long nights .
On the other hand, some controlled trials of melatonin have observed no suppressive effects on reproductive hormones . In a placebo-controlled crossover trial for one month (spring) and 3 weeks (fall), 2 mg melatonin given at 17:00 had no effect on LH . Melatonin 1.5 mg administered at 1600 for 8 days had no effect . Melatonin 10 mg administered at 10 PM had no effect on LH or FSH over 28–40 days . Morning acute administration of 3 mg melatonin was found to amplify LH pulses and area under the curve almost 50% among women in the early follicular phase .
In menopausal women, hot flashes are often synchronous with pulsatile release of LH [23, 24]. Although hot flashes may not be directly caused by either GnRH or LH, the same hypothalamic pulse generator may initiate both the hot flash and the pulsatile endocrine release. Thus, if melatonin suppressed this pulse generator in menopausal women, it might moderate hot flashes. In our group, Suhner et al. reasoned that by suppressing hot flashes, melatonin might be an effective treatment for menopausal sleep disturbances . In four subjects, they found that 3 mg melatonin at bedtime significantly decreased nocturnal LH secretion after 2 weeks. However, there was no significant reduction in hot flashes. Likewise, in another study of melatonin effects on LH, there was no significant reduction of hot flashes .
The current study was planned as an expanded test of the hypothesis that melatonin could suppress LH and relieve hot flashes by suppression of the hypothalamic pulse generator. Wishing to mimic the broader nocturnal melatonin elevation occurring naturally in winter, we gave melatonin not at bedtime but either 2.5–3.0 hours before bedtime or at the time of awakening. We intended to compare effects of evening and morning melatonin administration, which might have opposite effects on circadian timing. Lewy et al. had suggested that excessive melatonin doses might be ineffective stimuli of the circadian system, because blood levels might be elevated for too long . Therefore, the relatively low oral dose of 0.5 mg was selected, considering that even this dose may produce a peak blood concentration well above the physiologic peak .
Postmenopausal women troubled by hot flashes were recruited for the research by newspaper advertising. Out of 34 women interviewed for the study, 20 signed written informed consent and completed the research. Almost all of the others withdrew before or shortly after signing consent, mostly because their symptoms were not suitable, because they did not wish to adhere to the protocol, or because inspection of arm veins suggested that establishment of an intravenous sampling line might be difficult. A steady dosage of hormone replacement therapy was not considered an exclusion if it did not eliminate hot flashes. The study adhered to the principles of the Helsinki Declaration and was approved and supervised by the local Institutional Review Boards.
During the baseline week and throughout the 4 weeks of randomized treatment, participants completed a Daily Log of Sleep and Hot Flashes. This consisted of daily notations of a) bedtime, b) wake-up time, c) the self-estimated percentage of time-in-bed asleep, d) number of hot flashes in 24 hours, and e) a 10 cm. visual analog estimate of daily hot flash severity ranging from "mildest" to "most severe." These self-observations were averaged on a weekly basis. In addition, at the end of each of the 5 weeks, the participants completed two scales for self-rating depression for the previous week, the Center for Epidemiological Studies Depression Scale (CESD) and the Quick Inventory of Depressive Symptomatology self-rating version (QIDS-SR), as well as the SAFTEE, a 96-item symptom inventory designed to monitor adverse effects in clinical trials .
At the end of the baseline week, participants were admitted to the General Clinical Research Center for 25 hours, so that venous blood could be sampled repeatedly. An intravenous catheter in a forearm, with long extensions, was established. So far as possible, 2.5 cc blood samples were drawn every 20 min. for 24 hours for assays of luteinizing hormone. Also at this first admission, the baseline CESD, QIDS-SR, and SAFTEE scales were collected, the participants were instructed in urine collection procedures, and the randomized coded medications were provided for 4 weeks of treatment at home. During the randomized treatment, participants collected a first morning urine specimen twice a week, measuring the volume, recording the collection intervals, and freezing aliquots, so that urinary LH excretion could be examined. On the last day of the 4 weeks of randomized treatment, participants were readmitted to the General Clinical Research Center for repeated 24-hour blood sampling and collection of all data forms. Participants who completed the study were paid $440, or pro-rata for early discontinuation.
The staff experienced considerable difficulty in maintaining the venous catheters and obtaining blood specimens. For the first 12 participants, of 1898 planned samples, only 1090 samples were successfully frozen. However, at least a few samples were obtained from each of the 12 participants, making it possible to estimate daily mean LH blood concentration. Because of the missed samples and the small number of subjects completing the blood sampling procedures, it was not possible to analyze the parameters of pulsatile LH secretion.
We had originally planned to complete study of 30 subjects to have reasonable power for the endocrine studies, recognizing that for the sleep, hot flashes, and symptom data, a much larger number of subjects would be desirable for adequate power . After 13 participants had completed the protocol (blood LH from one could not be used), it was unfortunately necessary to discontinue the inpatient studies, because of loss of GCRC funding. An additional 7 subjects completed the protocol entirely at home, collecting urine but no blood samples. Resources did not permit recruitment of the final 10 subjects planned.
LH levels in blood and urine were measured using DSL-10-460 ActiveR LH Elisa kits (Diagnostic System Laboratories, Inc, Webster, TX.) This is an enzymatically amplified "one-step" sandwich-type immunoassay. The manufacturer states that there is no detectable cross-reactivity with TSH, FSH, and hCG. Blood samples were measured by direct addition of 50 μL serum, whereas urine samples were measured by diluting 1:1 with zero standards (each at 25 μL). For accuracy, above-scale LH urine samples (> 180 mIU/ml) were re-assayed at increased dilutions. The sensitivity of the EIA is 0.1 mIU/ml. At mean LH concentrations of 3.2 mIU/ml and 30.8 mIU/ml, intra-assay coefficients of variation were 2.2% and 4.9% respectively, and inter-assay CVs were 2.9% and 4.8% respectively. Urinary rates of LH excretion (mIU/hr) were computed by multiplying the urine LH concentration (mIU/ml) by the urine excretion rate (ml/hr) calculated over the interval between voidings. Measured urinary LH excretion was quite erratic from day to day, with the daily standard deviation being 69% of the mean. Therefore, for each subject, daily urine LH mIU/hr values outside a range of ± 2SD were ignored, in most cases leaving 8–9 values for computation of slopes. The first and last values meeting the ± 2SD criteria were also selected to explore changes in urinary LH excretion during treatment.
Contrasts between groups were examined using ANOVA, with the initial score as the covariate. When appropriate due to the distributions, these results were confirmed with nonparametric analyses, which in the case of LH measures, were applied to change scores.
The 20 participants completing the study had a mean age of 59 years (range 51–69).
The rank-order correlation of average blood LH at baseline versus the first urine excretion sample was rS = 0.84 (P < 0.001). The final treatment measures were correlated rS = 0.73 (P < 0.01). The adjusted changes in urine LH excretion and blood LH concentration were positively but not significantly correlated.
Reported hot flashes per day averaged 8.9 in the baseline week and 7.8 in the final week of randomized treatment. There was no significant difference between groups in hot flashes change, but the trend was for the placebo group to report the most improvement (contrary to prediction). Similarly, reported severity on the visual-analog scales dropped from 47 at baseline to 37 in the final week, with the greatest drop in the placebo group, though group contrasts were not significant. Likewise, changes were not significant between groups in the product of the reported number of hot flashes times the reported severity. Changes in hot flashes from baseline to the end of treatment were not significantly correlated with changes in blood or urinary LH over the same interval.
In the baseline week, the participants reported an average bedtime of 10:52 PM and an average uptime of 06:41 AM, with an average of 85% of the time-in-bed spent asleep. Though reported sleep percent increased to 88% at the end of treatment, there were no significant contrasts between treatment groups. The prediction that evening melatonin administration would phase-shift sleep times earlier than would morning melatonin (based on the melatonin phase-response curve)  was not confirmed.
On the CESD, mood improved from an average score of 7.5 at baseline to 6.9 in the final week. A CESD score >16, which is a sensitive screening criterion for major depression, was reported by 4 women at baseline, but only by 2 at the end of treatment. Perhaps because the group which would receive morning melatonin included the participant with the highest CESD (which improved during the study), that group showed the greatest improvement, but the contrasts between groups were not significant. Similarly, the average QIDS scores improved from 5.2 to 4.3 from baseline to the end of randomized treatments, with no significant differences between groups.
There were no clinically significant or severe adverse effects reported during randomized treatment with melatonin or placebo. Some women felt that the medication caused mild sleepiness, but it was tolerable. The SAFTEE symptom inventory displayed a 9.2% average reduction of diverse medical symptom reports from baseline to the end of treatment. Seventeen classes of symptoms and 10 individual items related to sleepiness and mood were examined, but there were no significant contrasts between randomized groups.
To summarize, this trial tended to confirm a preliminary study of Suhner et al.  and support the hypothesis that melatonin given to postmenopausal women can suppress luteinizing hormone as compared to placebo. This result was statistically marginal. The amount of suppression compared to placebo was very modest, considering the high levels of secretion among women of this age. Indeed, since participants receiving placebo showed higher LH at the end of treatment than at baseline, suppression in the group receiving melatonin in the evening was suggested merely by a smaller percentage increase than in the placebo group. The directions of change were similar in both blood and urine LH. Urine results were far more erratic, but perhaps because data for urine LH were available for more subjects, the contrasts were somewhat more statistically significant. Because fewer subjects than planned were studied and the statistical outcome was not conclusive, these results certainly need replication.
It was interesting to observe that morning melatonin was associated with more suppression of LH than evening melatonin, though not significantly so. Perhaps melatonin administration 4–5 hours before bedtime might have served to make the evening onset of melatonin more dramatically early and more effective. The investigators had not chosen to give evening melatonin 4–5 hours before bedtime, because it was believed that evening sedation and a possibly unwanted advance in sleep timing might be more severe with earlier administration.
It would be important to examine melatonin effects on the reproductive endocrinology of women of other ages and effects on men. Due to the excess LH secretion of postmenopausal women, effects of melatonin were not expected to be harmful to this group. However, if melatonin interferes with the reproductive endocrinology of women attempting to have children or suppresses LH (and thus testosterone) among men, such effects might be quite unwelcome. Since melatonin is becoming popular in the treatment of children with developmental problems, it would also be important to explore whether melatonin interferes with their endocrine development.
The observations were consistent with the hypothesis that photoperiodic mechanisms are active in human physiology. The timing chosen (well before bedtime or after arising) and the low dosage of melatonin, designed to simulate effects of a natural winter photoperiod upon the melatonin profile, may explain why we were able to demonstrate this marginal melatonin effect on LH, when other studies have not had similar results. Both the before-bedtime and morning administration would tend to extend the duration of the nightly melatonin elevation. Possibly, other studies have failed to document an effect of melatonin on LH because bedtime dosing or use of higher doses creates 24-hour melatonin profiles less similar to those produced by a natural winter photoperiod.
Nevertheless, it might be possible to further optimize timing and dosage.
The results of this study did not support our hypothesis that melatonin would improve sleep and suppress hot flashes among symptomatic postmenopausal women, but there were no significant adverse effects, and LH was possibly suppressed. There were too few subjects to provide adequate power for exploration of symptomatic effects, especially considering the well-known placebo responses in studies of this type . Therefore, larger studies of melatonin effects on postmenopausal women with hot flashes are needed.
Supported by Scripps Clinic Academic Affairs and by the Scripps Clinic General Clinical Research Center (NIH M01 RR00833). Assays were supported by NIMH grant to UCSD MH68545. The hard work of the participants and the contribution of the GCRC nurses, especially Robin Morrisey, R.N., Janine Kampelman, R.N., and Nichole Bianchi Mareno, R.N. are appreciated. James A. Koziol, Ph.D. was the statistical consultant. Marianne Lewis, R.Ph., Investigational Drug Pharmacist, prepared the coded research medications.
- Ferrari E, Fraschini F, Brambilla F: Hormonal circadian rhythms in eating disorders. Biol Psychiatry. 1990, 27: 1007-1020. 10.1016/0006-3223(90)90037-3.View ArticlePubMedGoogle Scholar
- Goldman BD: Mammalian photoperiodic system: Formal properties and neuroendocrine mechanisms of photoperiodic time measurement. J Biol Rhythms. 2001, 16: 283-301. 10.1177/074873001129001980.View ArticlePubMedGoogle Scholar
- Anand S, Losee-Olson S, Turek FW, Horton TH: Differential regulation of luteinizing hormone and follicle-stimulating hormone in male Siberian hamsters by exposure to females and photoperiod. Endocrinology. 2002, 143: 2178-2188. 10.1210/en.143.6.2178.View ArticlePubMedGoogle Scholar
- Lincoln GA: Neuroendocrine regulation of seasonal gonadotrophin and prolactin rhythms: lessons from the Soay ram model. Reprod Suppl. 2002, 59: 131-147.PubMedGoogle Scholar
- Cohen M, van Heusden AM, Verdonk HER, Wijnhamer P: Melatonin/Norethisterone contraception. Melatonin and the Pineal Gland- From Basic Science to Clinical Application. Edited by: Touitou Y, Arendt J and Pevet P. 1993, Amsterdam, Elsevier Science Publishers, 339-345.Google Scholar
- Bellipanni G, Bianchi P, Pierpaoli W, Bulian D, Ilyia E: Effects of melatonin in perimenopausal and menopausal women: a randomized and placebo controlled study. Experimental Gerontology. 2001, 36: 297-310. 10.1016/S0531-5565(00)00217-5.View ArticlePubMedGoogle Scholar
- Kadva A, Djahanbakhch O, Monson J, Di WL, Silman R: Elevated nocturnal melatonin is a consequence of gonadotropin-releasing hormone deficiency in women with hypothalamic amenorrhea. J Clin Endocrinol Metab. 1998, 83: 3653-3662. 10.1210/jc.83.10.3653.View ArticlePubMedGoogle Scholar
- Laughlin GA, Loucks AB, Yen SSC: Marked augmentation of nocturnal melatonin secretion in amenorrheic athletes, but not in cycling athletes: unaltered by opioidergic or dopaminergic blockade. J Clin Endocrinol Metab. 1991, 73 (6): 1321-6.View ArticlePubMedGoogle Scholar
- Luboshitzky R, Wagner O, Lavi S, Herer P, Lavie P: Abnormal melatonin secretion in male patients with hypogonadism. J Mol Neurosci. 1996, 7 (2): 91-8.View ArticlePubMedGoogle Scholar
- Luboshitzky R, Qupti G, Ishay A, Shen-Orr Z, Futerman B, Linn S: Increased 6-sulfatoxymelatonin excretion in women with polycystic ovary syndrome. Fertil Steril. 2001, 76: 506-510. 10.1016/S0015-0282(01)01930-6.View ArticlePubMedGoogle Scholar
- Berga SL, Mortola JF, Yen SSC: Amplification of nocturnal melatonin secretion in women with functional hypothalamic amenorrhea. J Clin Endocrinol Metab. 1988, 66: 242-244.View ArticlePubMedGoogle Scholar
- Piovesan A, Terzolo M, Revelli A, Guidetti D, Puligheddu B, Codegone A, Ali A, Cassoni P, Massobrio M, Angeli A: Sex-related effects of exogenous melatonin on the pulsatile secretion of prolactin, TSH and LH in healthy human subjects. Melatonin and the Pineal Gland - From Basic Science to Clinical Application. Edited by: Touitou Y, Arendt J and Pevet P. 1993, Amsterdam, Elsevier Science Publishers, 355-358.Google Scholar
- Luboshitzky R, Shen-Orr Z, Shochat T, Herer P, Lavie P: Melatonin administered in the afternoon decreases next-day luteinizing hormone levels in men. Journal of Molecular Neuroscience. 1999, 75-80. 10.1385/JMN:12:1:75.Google Scholar
- Luboshitzky R, Shen-Orr Z, Nave R, Lavi S, Lavie P: Melatonin administration alters semen quality in healthy men. J Androl. 2002, 23: 572-578.PubMedGoogle Scholar
- Wehr TA: Effect of seasonal charges in daylength on human neuroendocrine function. Horm Res. 1998, 49: 118-124. 10.1159/000023157.View ArticlePubMedGoogle Scholar
- Wehr TA: The durations of human melatonin secretion and sleep respond to changes in daylength (photoperiod). J Clin Endocrinol Metab. 1991, 73: 1276-1280.View ArticlePubMedGoogle Scholar
- Wehr TA, Moul DE, Barbato G, Giesen HA, Seidel JA, Barker C, Bender C: Conservation of photoperiod-responsive mechanisms in humans. Amer J Physiol. 1993, 265: R846-57.PubMedGoogle Scholar
- Luboshitzky R, Levi M, Shen-Orr Z, Blumenfeld Z, Herer P, Lavie P: Long-term melatonin administration does not alter pituitary-gonadal hormone secretion in normal men. Hum Repro. 2000, 15: 60-65. 10.1093/humrep/15.1.60.View ArticleGoogle Scholar
- Wright J, Aldhous M, Franey C, English J, Arendt J: The effects of exogenous melatonin on endocrine function in man. Clin Endocrinol (Oxf). 1986, 24: 375-382.View ArticleGoogle Scholar
- Rajaratnam SMW, Dijk DJ, Middleton B, Stone BM, Arendt J: Melatonin phase-shifts human circadian rhythms with no evidence of changes in the duration of endogenous melatonin secretion or the 24-hour production of reproductive hormones. The Journal of Clinical Endocrinology & Metabolism. 2003, 88: 4303-4309. 10.1210/jc.2003-030460.View ArticleGoogle Scholar
- Seabra MLV, Bignotto M, Pinto LRJ, Tufik S: Randomized, double-blind clinical trial, controlled with placebo, of the toxicology of chronic melatonin treatment. J Pineal Res. 2000, 29: 193-200. 10.1034/j.1600-0633.2002.290401.x.View ArticlePubMedGoogle Scholar
- Cagnacci A, Elliott JA, Yen SSC: Amplification of pulsatile LH secretion by exogenous melatonin in women. J Clin Endocrinol Metab. 1991, 73: 210-212.View ArticlePubMedGoogle Scholar
- Casper RF, Yen SSC: Neuroendocrinology of menopausal flushes: an hypothesis of flush mechanism. Clin Endocrinol. 1985, 22: 293-312.View ArticleGoogle Scholar
- Kronenberg F, Downey JA: Thermoregulatory physiology of menopausal hot flashes: a review. Can J Physiol Pharmacol. 1987, 65: 1312-1324.View ArticlePubMedGoogle Scholar
- Suhner AG, Poceta JS, Kline LE, Murphy PJ, Campbell SS: Melatonin as treatment for sleep disturbances during menopause. Sleep. 2003, 26: A160.Google Scholar
- Lewy AJ, Emens JS, Sack RL, Hasler BP, Bernert RA: Low, but not high, doses of melatonin entrained a free-running blind person with a long circadian period. Chronobiol Intl. 2002, 19: 649-658. 10.1081/CBI-120004546.View ArticleGoogle Scholar
- Levine J: Standardized techniques for side-effect assessment: early result using SAFTEE. Psychopharmacol Bull. 1987, 23: 91-105.Google Scholar
- Sloan JA, Loprinzi CL, Novotny PJ, Barton DL, Lavasseur BI, Windschitl H: Methodologic lessons learned from hot flash studies. J Clin Oncol. 2001, 19: 4280-4290.PubMedGoogle Scholar
- Lewy AJ, Bauer VK, Ahmed S, Thomas KH, Cutler NL, Singer CM, Moffit MT, Sack RL: The human phase response curve (PRC) to melatonin is about 12 hours out of phase with the PRC to light. Chronobiol Intl. 1998, 15 (1): 71-83.View ArticleGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1472-6874/6/8/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.